What an unfortunate thread...
It seems that we get people like this two or three times a year. The details vary, but the gist is that they want to deconstruct what someone else has done in order to make themselves look important/intelligent/non-gullible. If you look through the threads here, you will find examples where people have tried (and sometimes succeeded) in picking fights with John Curl, Charles Hansen, Nelson Pass, et. al.
To me, the entire thread became clear when the original poster said something to the effect that a 'long tailed pair' is different from a differential. (My response to such a preposterous claim would be rendered in rows of *'s, so I will refrain from typing words that will only be edited.) In other words, he's trying to go the semantics route. Gee, like we haven't had ten or twenty threads like that this year already. The bright side of 'semantics' threads is that you can quickly and efficiently find out which side of the fence any given poster is on. The dark side is that it seems to bring out the worst in many posters; boorish, rude, arrogant, and obnoxious behavior predominate.
I am pleased to note that several members have pleaded for sanity in this thread; a much higher than normal percentage, in fact. Including one (gasp! dare I say it?) moderator.
Sadly, there's no patent office that would grant a patent on poor behavior. If there were, I would apply for the patent, then charge such an exorbitant fee to license the concept that no one would be able to afford the useage thereof...regardless of whether it was obvious or non-obvious.
Grey
P.S.: Having recently been sucked into a thread where someone persisted in misreading my posts--whether accidentally or on purpose--I find it especially unpleasant to find the same behavior in this thread. It's annoying, to say the least.
By the way, when Nelson suggests that you build a circuit to watch it in action before making more yak-yak...you had best go build the damned thing.
Or, in what is more-or-less my motto around these parts: Reality beats theory seven days a week.
It seems that we get people like this two or three times a year. The details vary, but the gist is that they want to deconstruct what someone else has done in order to make themselves look important/intelligent/non-gullible. If you look through the threads here, you will find examples where people have tried (and sometimes succeeded) in picking fights with John Curl, Charles Hansen, Nelson Pass, et. al.
To me, the entire thread became clear when the original poster said something to the effect that a 'long tailed pair' is different from a differential. (My response to such a preposterous claim would be rendered in rows of *'s, so I will refrain from typing words that will only be edited.) In other words, he's trying to go the semantics route. Gee, like we haven't had ten or twenty threads like that this year already. The bright side of 'semantics' threads is that you can quickly and efficiently find out which side of the fence any given poster is on. The dark side is that it seems to bring out the worst in many posters; boorish, rude, arrogant, and obnoxious behavior predominate.
I am pleased to note that several members have pleaded for sanity in this thread; a much higher than normal percentage, in fact. Including one (gasp! dare I say it?) moderator.
Sadly, there's no patent office that would grant a patent on poor behavior. If there were, I would apply for the patent, then charge such an exorbitant fee to license the concept that no one would be able to afford the useage thereof...regardless of whether it was obvious or non-obvious.
Grey
P.S.: Having recently been sucked into a thread where someone persisted in misreading my posts--whether accidentally or on purpose--I find it especially unpleasant to find the same behavior in this thread. It's annoying, to say the least.
By the way, when Nelson suggests that you build a circuit to watch it in action before making more yak-yak...you had best go build the damned thing.
Or, in what is more-or-less my motto around these parts: Reality beats theory seven days a week.
(Although I'm not here anymore...)
Yes, it seems I goofed. Or something.
Nelson had said:-
And I'd responded:-
I'd thought he was referring to just an uncoupled pair of amplifiers (as not being what his 'invention' was). (If you read his "DISCLOSURE OF THE INVENTION" in his patent, you'll see why I thought that that was what he meant. In that patent, it's clear that it's the coupling of the two (otherwise separate) amplifiers that's supposed to be the key, significant thing.)
But now I realise that wasn't what he meant. I now believe that what he meant was that it wasn't just an ordinary long-tailed pair ("differential pair", whatever), but that it was one with negative feedback applied, on both sides!
Shame on me for actually reading his patent. Shame on me for thinking he meant something less obvious than applying feedback to an amplifier.
GRollins said:
Yes, it seems I goofed. Or something.
Nelson had said:-
And I'd responded:-
I'd thought he was referring to just an uncoupled pair of amplifiers (as not being what his 'invention' was). (If you read his "DISCLOSURE OF THE INVENTION" in his patent, you'll see why I thought that that was what he meant. In that patent, it's clear that it's the coupling of the two (otherwise separate) amplifiers that's supposed to be the key, significant thing.)
But now I realise that wasn't what he meant. I now believe that what he meant was that it wasn't just an ordinary long-tailed pair ("differential pair", whatever), but that it was one with negative feedback applied, on both sides!
Shame on me for actually reading his patent. Shame on me for thinking he meant something less obvious than applying feedback to an amplifier.
Sigh. You made me go read the patent and I'm convinced that your argument holds no water. I'm certainly no expert in solid state design, but I do know what a long tailed pair is. This is no simple long tailed pair- the key, pointed out quite explicitly in the spec and claims, is indeed the cross coupling between the halves. If you wanted to moan about prior art and obviousness, van Scoyoc would be more relevant (but still not relevant).
"I can give you an answer, but I cannot give you an understanding." Samuel Johnson was always a pithy guy. If you want to read the file wrapper, perhaps you'll get a better understanding at least of the examiners' position.
"I can give you an answer, but I cannot give you an understanding." Samuel Johnson was always a pithy guy. If you want to read the file wrapper, perhaps you'll get a better understanding at least of the examiners' position.
Okay, I'm back
(I did just draw a nice, little diagram of an amplifier that fits the description in Claim 1 of the patent. Alas, I can't post it )
Take a look at the schematic on the Wikipedia article on differential amplifiers.
1. Split Re into two resistors in parallel, each with resistance Re/2.
2. On the left-hand side of the amplifier, connect the output to the input by a feedback resistor.
3. Attach an input resistor to the input.
4. Apply steps 2 and 3 to the right-hand side, as well.
Now see if the resulting amplifier fits the description in Claim 1.
The change in step 1 was utterly trivial. The connection at the top of the parallel pair of resistors (the connection of the emitters to each other) is the cross-coupling, the "means for coupling said first and second coupling terminals together", which is the key, essential ingredient in the "DISCLOSURE OF THE INVENTION". But that connection - between the emitters - existed before we split Re into two. The remaining steps just added negative feedback.
We could try other variations, such as having a constant current source in place of Re (a truly long-tailed pair, no?), and splitting that in two (maintaining the connection at the top of the current source(s)). But we've still got the essential cross-coupling (before and after the trivial splitting), so it's not a relevant difference. We could put a resistor in the cross-coupling (and we could even cheekily set the resistance to zero ), to fit with the "DISCLOSURE OF THE INVENTION", but even with zero resistance, the cross-coupling still plays the same, essential role as in the "DISCLOSURE". We could put folded cascodes in the outputs, replace the BJTs with MOSFETs or triodes, we could put resistors between the emitters (sources, cathodes, whatever) and the cross-coupling, or we could completely omit such changes, and still the cross-coupling plays the same, essential, vital, key role that it does in the "DISCLOSURE". They are not relevant differences.
In the patent, the key, essential thing is the cross-coupling. But if we start off with something that already has that cross-coupling (such as a long-tailed pair), then the key thing is the application of negative feedback. It's essentially the same thing, just looked at from different directions. It doesn't matter how non-obvious it might have been from one direction (as in the patent), it was still, screamingly obvious from the other (starting off with a long-tailed pair (or similar), and - surprise, surprise! - adding feedback).
SY said:
(I did just draw a nice, little diagram of an amplifier that fits the description in Claim 1 of the patent. Alas, I can't post it
)Take a look at the schematic on the Wikipedia article on differential amplifiers.
1. Split Re into two resistors in parallel, each with resistance Re/2.
2. On the left-hand side of the amplifier, connect the output to the input by a feedback resistor.
3. Attach an input resistor to the input.
4. Apply steps 2 and 3 to the right-hand side, as well.
Now see if the resulting amplifier fits the description in Claim 1.
The change in step 1 was utterly trivial. The connection at the top of the parallel pair of resistors (the connection of the emitters to each other) is the cross-coupling, the "means for coupling said first and second coupling terminals together", which is the key, essential ingredient in the "DISCLOSURE OF THE INVENTION". But that connection - between the emitters - existed before we split Re into two. The remaining steps just added negative feedback.
We could try other variations, such as having a constant current source in place of Re (a truly long-tailed pair, no?), and splitting that in two (maintaining the connection at the top of the current source(s)). But we've still got the essential cross-coupling (before and after the trivial splitting), so it's not a relevant difference. We could put a resistor in the cross-coupling (and we could even cheekily set the resistance to zero
), to fit with the "DISCLOSURE OF THE INVENTION", but even with zero resistance, the cross-coupling still plays the same, essential role as in the "DISCLOSURE". We could put folded cascodes in the outputs, replace the BJTs with MOSFETs or triodes, we could put resistors between the emitters (sources, cathodes, whatever) and the cross-coupling, or we could completely omit such changes, and still the cross-coupling plays the same, essential, vital, key role that it does in the "DISCLOSURE". They are not relevant differences.In the patent, the key, essential thing is the cross-coupling. But if we start off with something that already has that cross-coupling (such as a long-tailed pair), then the key thing is the application of negative feedback. It's essentially the same thing, just looked at from different directions. It doesn't matter how non-obvious it might have been from one direction (as in the patent), it was still, screamingly obvious from the other (starting off with a long-tailed pair (or similar), and - surprise, surprise! - adding feedback).
Sometimes the patent doesn't anticipate all the places a
concept might go, and sometimes a wise patent attorney
limits the discussion so as to get the patent through with a
minimum of fuss.
Black's patent on negative feedback is a fine case in point.
Reading it, you get a sense that he had no idea how far the
concept would go, and to what universal application. The X
patent is just another example, but the patent office saw it
as sufficiently unique and sufficiently narrow, and spit it right
on through.
You can't deny a patent on a paper clip because wire and the
bending of wire and the idea of fastening paper sheets together
are all obvious. When the wire is bent in a particular way as to
fasten paper, and no one else has bent it that way before, you
could probably obtain a patent.
concept might go, and sometimes a wise patent attorney
limits the discussion so as to get the patent through with a
minimum of fuss.
Black's patent on negative feedback is a fine case in point.
Reading it, you get a sense that he had no idea how far the
concept would go, and to what universal application. The X
patent is just another example, but the patent office saw it
as sufficiently unique and sufficiently narrow, and spit it right
on through.
You can't deny a patent on a paper clip because wire and the
bending of wire and the idea of fastening paper sheets together
are all obvious. When the wire is bent in a particular way as to
fasten paper, and no one else has bent it that way before, you
could probably obtain a patent.
Things are clearing up a little bit now. It seems that when Simon talks about crosscoupling, he refers to the "bridge resistor" between the emitters/drains. When I said crosscoupling I referred to the actual feedback from the outputs crossing between halves, as I have seen been done before in SUSY implementations. However, now when I look back I see a figure from the patent posted in this thread, and that figure does not have the type of crosscoupling I meant, so I suppose that is a later modification that was not part of the original patent.
Regarding the circuit in the figure, the two input transistors obviously form a circuit that is functionally equivalent to a diff pair, and which others use and have used. Is it a diff pair? Well, it is trivially functionally equivalent to one, but so is a Rush cascode, which is usually called a Rush cascode, not a diff pair. However, this circuit which is functionally equivalent to a diff pair is in my understanding used in a non-conventional way, so it is not this subcircuit alone that is patented, but the whole circuit and the way of using it, that is, as a balanced amplifier. Some would say that is straightforward and trivial, perhaps, but that is not important for a patent. If nobody has used such a circuit for the same puropse before, then it is patentable. On the other hand, if somebody has done exactly the same thing before but used a more traditional-looking diff pair, with one tail, then it is a hairier issue. The invention would then be functionally equivalent to prior art and used for the same thing. This would end up with a court ruling whether delta-triangle transformations are trivial or not.
This does however raise certain interesting questions on patentability in general. Can a circuit that is functionally equivlent to prior art be patented? That is, does one patent the actual toplogy of connection of the components, or does one patent the functional behaviour? There is probably no clear rule here. What happens if I make a copy of Nelsons circuit but use two resistors in series instead of the single bridge resistor? Can I claim it is a different circuit although it is clearly functionally equivalent? Probably this modification is so trivial that it would not be accepted, but then it is all a question of deciding what variations are trivial and not. Suppose I claim that using two resistors, coupled in opposite direction (some people claim to hear such things, you know) makes the circuit behave differently and better, could I then get a patent on this modification? Do I need to somehow prove that it makes a difference? If so, why? Nelson didn't have to prove anything about his circuit, did he? My guess is that it is next to impossible to figure out in advance what is patentable and what isn't, and even harder to guess what conclusion a court would come to. I have heard of a case where a US court ruled that a patent was still valid even though the invention was already published in very well known textbooks. If that is not prior art, what is? I have also myself read very carefully one swedish patent where the major claim about the invention is simply plainly wrong. I does not do what they claim, and if it did, it would be an understatement to say it would cause a sensation. I wonder what would happen if I made a copy of that invention but didn't claim it to do what it anyway doesn't do?
Regarding the circuit in the figure, the two input transistors obviously form a circuit that is functionally equivalent to a diff pair, and which others use and have used. Is it a diff pair? Well, it is trivially functionally equivalent to one, but so is a Rush cascode, which is usually called a Rush cascode, not a diff pair. However, this circuit which is functionally equivalent to a diff pair is in my understanding used in a non-conventional way, so it is not this subcircuit alone that is patented, but the whole circuit and the way of using it, that is, as a balanced amplifier. Some would say that is straightforward and trivial, perhaps, but that is not important for a patent. If nobody has used such a circuit for the same puropse before, then it is patentable. On the other hand, if somebody has done exactly the same thing before but used a more traditional-looking diff pair, with one tail, then it is a hairier issue. The invention would then be functionally equivalent to prior art and used for the same thing. This would end up with a court ruling whether delta-triangle transformations are trivial or not.
This does however raise certain interesting questions on patentability in general. Can a circuit that is functionally equivlent to prior art be patented? That is, does one patent the actual toplogy of connection of the components, or does one patent the functional behaviour? There is probably no clear rule here. What happens if I make a copy of Nelsons circuit but use two resistors in series instead of the single bridge resistor? Can I claim it is a different circuit although it is clearly functionally equivalent? Probably this modification is so trivial that it would not be accepted, but then it is all a question of deciding what variations are trivial and not. Suppose I claim that using two resistors, coupled in opposite direction (some people claim to hear such things, you know) makes the circuit behave differently and better, could I then get a patent on this modification? Do I need to somehow prove that it makes a difference? If so, why? Nelson didn't have to prove anything about his circuit, did he? My guess is that it is next to impossible to figure out in advance what is patentable and what isn't, and even harder to guess what conclusion a court would come to. I have heard of a case where a US court ruled that a patent was still valid even though the invention was already published in very well known textbooks. If that is not prior art, what is? I have also myself read very carefully one swedish patent where the major claim about the invention is simply plainly wrong. I does not do what they claim, and if it did, it would be an understatement to say it would cause a sensation. I wonder what would happen if I made a copy of that invention but didn't claim it to do what it anyway doesn't do?
Here's a fun thing to try
In the patent, it's the cross-coupling resistor that's the key, central thing. On the schematic on the patent cover page, it's resistor 40.
(While we're at it, let's also note that resistors 42 and 43 are connected to ground.)
Connect the middle of the cross-coupling resistor (resistor 40 in the patent cover page schematic) to ground. (Easily done when that resistor consists of two, equal resistors in series.)
In the patent, it's the cross-coupling resistor that's the key, central thing. On the schematic on the patent cover page, it's resistor 40.
(While we're at it, let's also note that resistors 42 and 43 are connected to ground.)
Connect the middle of the cross-coupling resistor (resistor 40 in the patent cover page schematic) to ground. (Easily done when that resistor consists of two, equal resistors in series.)
Nelson Pass said:
About now one finds it impossible to avoid pointing any interested person at Henry Petroski's work. In particular The evolution of useful things, The Pencil and To engineer is human. If you are involved in any way in any of the engineering arts you can't consider yourself a well rounded human being unless you have read these.
Henry spends a great deal of time looking at the paperclip in The evolution of useful things. If you want to find out what it means to invent this is a very good place to start.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.